File size: 9,278 Bytes
ca3e86b
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
import argparse
import os

import gradio as gr
from text_generation import Client

TITLE = """<h2 align="center">🚀 Falcon-Chat demo</h2>"""
USER_NAME = "User"
BOT_NAME = "Falcon"
DEFAULT_INSTRUCTIONS = f"""The following is a conversation between a highly knowledgeable and intelligent AI assistant, called Falcon, and a human user, called User. In the following interactions, User and Falcon will converse in natural language, and Falcon will answer User's questions. Falcon was built to be respectful, polite and inclusive. Falcon was built by the Technology Innovation Institute in Abu Dhabi. Falcon will never decline to answer a question, and always attempts to give an answer that User would be satisfied with. It knows a lot, and always tells the truth. The conversation begins.
"""
RETRY_COMMAND = "/retry"
STOP_STR = f"\n{USER_NAME}:"
STOP_SUSPECT_LIST = [":", "\n", "User"]

INFERENCE_ENDPOINT = os.environ.get("INFERENCE_ENDPOINT")
INFERENCE_AUTH = os.environ.get("INFERENCE_AUTH")


def chat_accordion():
    with gr.Accordion("Parameters", open=False):
        temperature = gr.Slider(
            minimum=0.1,
            maximum=2.0,
            value=0.8,
            step=0.1,
            interactive=True,
            label="Temperature",
        )
        top_p = gr.Slider(
            minimum=0.1,
            maximum=0.99,
            value=0.9,
            step=0.01,
            interactive=True,
            label="p (nucleus sampling)",
        )
    return temperature, top_p


def format_chat_prompt(message: str, chat_history, instructions: str) -> str:
    instructions = instructions.strip(" ").strip("\n")
    prompt = instructions
    for turn in chat_history:
        user_message, bot_message = turn
        prompt = f"{prompt}\n{USER_NAME}: {user_message}\n{BOT_NAME}: {bot_message}"
    prompt = f"{prompt}\n{USER_NAME}: {message}\n{BOT_NAME}:"
    return prompt


def chat(client: Client):
    with gr.Column(elem_id="chat_container"):
        with gr.Row():
            chatbot = gr.Chatbot(elem_id="chatbot")
        with gr.Row():
            inputs = gr.Textbox(
                placeholder=f"Hello {BOT_NAME} !!",
                label="Type an input and press Enter",
                max_lines=3,
            )

    with gr.Row(elem_id="button_container"):
        with gr.Column():
            retry_button = gr.Button("♻️ Retry last turn")
        with gr.Column():
            delete_turn_button = gr.Button("🧽 Delete last turn")
        with gr.Column():
            clear_chat_button = gr.Button("✨ Delete all history")

    gr.Examples(
        [
            ["Hey Falcon! Any recommendations for my holidays in Abu Dhabi?"],
            ["What's the Everett interpretation of quantum mechanics?"],
            ["Give me a list of the top 10 dive sites you would recommend around the world."],
            ["Can you tell me more about deep-water soloing?"],
            ["Can you write a short tweet about the Apache 2.0 release of our latest AI model, Falcon LLM?"],
        ],
        inputs=inputs,
        label="Click on any example and press Enter in the input textbox!",
    )

    with gr.Row(elem_id="param_container"):
        with gr.Column():
            temperature, top_p = chat_accordion()
        with gr.Column():
            with gr.Accordion("Instructions", open=False):
                instructions = gr.Textbox(
                    placeholder="LLM instructions",
                    value=DEFAULT_INSTRUCTIONS,
                    lines=10,
                    interactive=True,
                    label="Instructions",
                    max_lines=16,
                    show_label=False,
                )

    def run_chat(message: str, chat_history, instructions: str, temperature: float, top_p: float):
        if not message or (message == RETRY_COMMAND and len(chat_history) == 0):
            yield chat_history
            return

        if message == RETRY_COMMAND and chat_history:
            prev_turn = chat_history.pop(-1)
            user_message, _ = prev_turn
            message = user_message

        prompt = format_chat_prompt(message, chat_history, instructions)
        chat_history = chat_history + [[message, ""]]
        stream = client.generate_stream(
            prompt,
            do_sample=True,
            max_new_tokens=1024,
            stop_sequences=[STOP_STR, "<|endoftext|>"],
            temperature=temperature,
            top_p=top_p,
        )
        acc_text = ""
        for idx, response in enumerate(stream):
            text_token = response.token.text

            if response.details:
                return

            if text_token in STOP_SUSPECT_LIST:
                acc_text += text_token
                continue

            if idx == 0 and text_token.startswith(" "):
                text_token = text_token[1:]

            acc_text += text_token
            last_turn = list(chat_history.pop(-1))
            last_turn[-1] += acc_text
            chat_history = chat_history + [last_turn]
            yield chat_history
            acc_text = ""

    def delete_last_turn(chat_history):
        if chat_history:
            chat_history.pop(-1)
        return {chatbot: gr.update(value=chat_history)}

    def run_retry(message: str, chat_history, instructions: str, temperature: float, top_p: float):
        yield from run_chat(RETRY_COMMAND, chat_history, instructions, temperature, top_p)

    def clear_chat():
        return []

    inputs.submit(
        run_chat,
        [inputs, chatbot, instructions, temperature, top_p],
        outputs=[chatbot],
        show_progress=False,
    )
    inputs.submit(lambda: "", inputs=None, outputs=inputs)
    delete_turn_button.click(delete_last_turn, inputs=[chatbot], outputs=[chatbot])
    retry_button.click(
        run_retry,
        [inputs, chatbot, instructions, temperature, top_p],
        outputs=[chatbot],
        show_progress=False,
    )
    clear_chat_button.click(clear_chat, [], chatbot)


def get_demo(client: Client):
    with gr.Blocks(
        # css=None
        # css="""#chat_container {width: 700px; margin-left: auto; margin-right: auto;}
        #        #button_container {width: 700px; margin-left: auto; margin-right: auto;}
        #        #param_container {width: 700px; margin-left: auto; margin-right: auto;}"""
        css="""#chatbot {
    font-size: 14px;
    min-height: 300px;
}"""
    ) as demo:
        gr.HTML(TITLE)

        with gr.Row():
            with gr.Column():
                gr.Image("home-banner.jpg", elem_id="banner-image", show_label=False)
            with gr.Column():
                gr.Markdown(
                    """**Chat with [Falcon-40B-Instruct](https://huggingface.co/tiiuae/falcon-40b-instruct), brainstorm ideas, discuss your holiday plans, and more!**

                    ✨ This demo is powered by [Falcon-40B](https://huggingface.co/tiiuae/falcon-40b), finetuned on the [Baize](https://github.com/project-baize/baize-chatbot) dataset, and running with [Text Generation Inference](https://github.com/huggingface/text-generation-inference). [Falcon-40B](https://huggingface.co/tiiuae/falcon-40b) is a state-of-the-art large language model built by the [Technology Innovation Institute](https://www.tii.ae) in Abu Dhabi. It is trained on 1 trillion tokens (including [RefinedWeb](https://huggingface.co/datasets/tiiuae/falcon-refinedweb)) and available under the Apache 2.0 license. It currently holds the 🥇 1st place on the [🤗 Open LLM leaderboard](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard). This demo is made available by the [HuggingFace H4 team](https://huggingface.co/HuggingFaceH4). 

                    🧪 This is only a **first experimental preview**: the [H4 team](https://huggingface.co/HuggingFaceH4) intends to provide increasingly capable versions of Falcon Chat in the future, based on improved datasets and RLHF/RLAIF.

                    👀 **Learn more about Falcon LLM:** [falconllm.tii.ae](https://falconllm.tii.ae/)

                    ➡️️ **Intended Use**: this demo is intended to showcase an early finetuning of [Falcon-40B](https://huggingface.co/tiiuae/falcon-40b), to illustrate the impact (and limitations) of finetuning on a dataset of conversations and instructions. We encourage the community to further build upon the base model, and to create even better instruct/chat versions!

                    ⚠️ **Limitations**: the model can and will produce factually incorrect information, hallucinating facts and actions. As it has not undergone any advanced tuning/alignment, it can produce problematic outputs, especially if prompted to do so. Finally, this demo is limited to a session length of about 1,000 words.
                    """
                )

        chat(client)

    return demo


if __name__ == "__main__":
    parser = argparse.ArgumentParser("Playground Demo")
    parser.add_argument(
        "--addr",
        type=str,
        required=False,
        default=INFERENCE_ENDPOINT,
    )
    args = parser.parse_args()
    client = Client(args.addr, headers={"Authorization": f"Basic {INFERENCE_AUTH}"})
    demo = get_demo(client)
    demo.queue(max_size=128, concurrency_count=16)
    demo.launch()